Usage of expensive features such as wildcard search, Boolean search, exposure of all refinements, or large numbers of records per page is a common cause of performance problems.
The Request Log Analyzer can be used to measure the performance of queries utilizing expensive features against "standard" queries. Because the Request Log Analyzer can segment performance numbers based on these features (see Extended Query Types and Request Profiling), it is trivial to compare statistics. However, beware that statistics from a single MDEX Engine necessarily are inter-dependent. A standard query may be stuck in the request queue behind an expensive query, thus inflating its response time. Similarly, a standard query may be executing on a thread simultaneously with an expensive query on a parallel thread. If the expensive query causes resource (CPU, disk) contention, the standard query will see an inflated engine time.